JP2002188955A - Strength deterioration detection method of structure by using ambiguous external force - Google Patents

Abstract

PROBLEM TO BE SOLVED: To estimate highly accurately the characteristic frequency of a structure by using an observed vibration waveform, and detect strength deterioration of the structure from irreversible fluctuation of the estimated characteristic frequency. SOLUTION: This method has a constitution by computer signal processing for determining the frequency at which a frequency distribution becomes maximum by using a means for dividing the observed vibration waveform with a short time to acquire partial waveforms, a means for approximating the partial waveforms with periodical waveforms, and a means for collecting the frequencies of the periodical waveforms to acquire the frequency distribution, and detecting the strength deterioration caused by elapse of time from the irreversible fluctuation of the frequency having the maximum distribution.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[Industrial applications] Used for nondestructive inspection in industrial fields such as civil engineering, architecture, ships, aircraft, and power plants.

[0002]

2. Description of the Related Art Visual observation is a conventionally known and widely practiced method for examining the state of strength deterioration and damage due to aging and earthquakes of structures and the progress of such deterioration. At the research level, there is a frequency response method that vibrates a structure with a periodic signal and compares the response with the past response, or an AE method (acoustic emission method) that observes ultrasonic waves emitted at the time of minute destruction of metal. .

[0003]

The frequency response method, which is one of the prior arts, usually requires a huge facility for exciting a structure, and the structure cannot be inspected at any time. There's a problem. In addition, the AE method mainly deals with a metal, and there is a problem that specialized knowledge is required in order to know the relationship between emitted sound waves and deterioration (metal fatigue). A further problem is that it is difficult to check during operation or use.

[0004] The problem to be solved by the present invention is to detect information relating to strength deterioration only from an observed vibration waveform of a structure vibrating in a normal use state without using special equipment for exciting the structure. Is to provide a way to

[0005]

In order to solve the above-mentioned problems, the present invention provides a means for obtaining a partial waveform by dividing an observed vibration waveform of a structure vibrating due to an uncertain external force at short intervals,
Means for approximating the partial waveform with a single periodic waveform, and means for determining the frequency distribution of the cycle of the periodic waveform approximating a series of the partial waveforms obtained from the observed vibration waveform, wherein the frequency distribution is maximized. Calculating the natural frequency of the structure from the cycle, and detecting the strength deterioration of the structure using an uncertain external force, wherein the strength deterioration of the structure is detected from the irreversible fluctuation of the natural frequency. The law is the means.

[0006]

The operation of the present invention will be described below with reference to equations and figures. In the band from the frequency f _L to f _H, uncertain external forces, such as the frequency components (power spectrum) can be regarded as approximately uniform with respect to (eg wind, earthquake, microtremor, vehicle traveling vibration) is driven by a structure that vibrates, The observed vibration waveform is represented by a time series W (t). However, if t = 1, 2,... And the sampling frequency is F hertz, W (t) is represented by data discretized at 1 / F second intervals.

Here, W (t) is changed from the first to the K-th partial waveform W _k (m) (k = 1,
2,..., K; m = 1, 2,..., M). W
_{Assuming that k} (m) is approximated by one periodic waveform S (m, T) having a period T,

(Equation 1) Where

(Equation 2) Then, S (m, T) gives a good approximation to W _k (m).

Assume that the frequency component of W (t) is limited to the band from f _P to f _Q ,

(Equation 3) , A period T _k that makes S (m, T) a good approximation of W _k (m) is

(Equation 4) Can be found in the range.

The above T _k can be determined with high accuracy by the following general harmonic analysis. Coefficients A (T) and B (T)

(Equation 5) And S (m, T) is given by

(Equation 6) And

Assuming that the residual waveform R (m) is given by (Equation 1), its power V (T) is

(Equation 7) And Then, by changing T, T (T) at which V (T) is minimized
And its value is set to T _k .

According to the above method, all the partial waveforms W
by obtaining a _{T k} from _k (m), from _{T k} is 1 / _{f Q} 1 /
to concentrate on during the f _P. FIG. 1 shows the spectrum of W (t).
1A, the frequency distribution P (T) of T _k is as shown in FIG. 1B assuming that K is sufficiently large.

[0012] Even if the frequency component of W (t) is restricted to the band from f _P (> f _L ) to f _Q (<f _H ),

(Equation 8) In the case of, a good approximation of W _k (m) is given by only one sinusoid S (m, T _k ) whose period is T _k obtained by the general harmonic analysis of (Equation 5) to (Equation 7). Can not.

[0013] In the case where the frequency band of W (t) is represented by (Equation 8), T _k is the first sine wave component obtained at the stage of sequentially detecting the spectrum of W _k (m) in the general harmonic analysis. Assuming that the spectrum of W (t) is shown in FIG. 2A, the frequency distribution P (t) of T _k assumes that K is sufficiently large as shown in FIG. Become.

The spectrum of W (t) is shown in FIGS.
As shown in FIG. 3A, which is the sum of (a),
The frequency distribution P (T) of T _k is as shown in FIG. 3B assuming that K is sufficiently large, and the bandwidth (f _Q ) of W (t) is obtained.
If -f _P) is smaller than F / M, the period P (T) is maximized _{T o} is the reciprocal 1 / _{f o} of the frequency located in the middle of _{f P} and _{f Q.}

Assuming that the natural frequency of the structure is f _o , the resonance sharpness is Q, and the resonance bandwidth is f _o / Q, the frequency of the observed vibration waveform W (t) of the structure is calculated by the operation of the present invention. Distribution P
(T), and if f _o / Q is smaller than F / M, P
T _o (= 1 / f _o ) is obtained from the maximum cycle of (T).

As a practical example, the natural frequency f _o of a building
3 Hz, the damping coefficient ratio h a (1 / 2Q) 0.04, 100 Hz the sampling frequency F of the observed vibration waveform, the data number M of partial waveforms When _40, f o _/Q=0.24, F
/M=2.5, which is 300 from the observed vibration waveform for 2 minutes.
It is possible to obtain the frequency distribution of the periods of points, it is possible to determine the T _o at the intervals of 0.01 seconds accuracy.

When the natural period of a structure fluctuates due to a temperature change during the day or night or a temperature change during the four seasons, the change is periodic and reversible. It becomes a fluctuation. Accordingly, the natural period of the structure is continuously measured by the above method, and the deterioration of the structure can be detected by detecting the irreversible fluctuation of the natural period.

If there is no band limitation on the observed vibration waveform W (t) and the frequency component outside the change range of T is not sufficiently small when finding the minimum value of V (T) in (Equation 7), P (T) shows a maximum at the upper and lower limits of T at both ends of the distribution P (T). So, both ends of the frequency distribution is assumed, except in the detection of T _o.

Before dividing into the partial waveform W _k (m), the maximum value generated at both ends of the frequency distribution P (T) can be suppressed by band-limiting the observed vibration waveform W (t). Such band limitation is performed as needed.

It should be noted that the effect of the present invention is that the influence of the spectrum of the uncertain external force on the detection of the natural frequency (or the natural period) is much smaller than the case where the detection is performed using the Fourier spectrum. . An example shows that the natural period can be detected with high accuracy even in a state where it can hardly be detected by the Fourier spectrum of the observed vibration waveform.

[0021]

FIG. 4 is a flow chart of computer software for _obtaining a natural frequency To of a structure from an observed vibration waveform W (t). 4 shows an example of a detection method. In the figure, reference numeral 1 denotes a partial waveform reading block which obtains a partial waveform W _k (m) by dividing the observed vibration waveform W (t) at short intervals, and 2 denotes a single periodic waveform for W _k (m). 3 is a waveform analysis block to be approximated, and 3 is a frequency distribution P of a cycle using the cycle _Tk obtained in the waveform analysis block.
(T) seeking a frequency distribution creation block, frequency distribution P the finished in the K period T _k (T) outputs the cycle becomes maximum as the natural period T _o of the structure.

In the embodiment shown in FIG. 4, 1 is calculated from K partial waveforms.
One of the obtained frequency distribution P (T), although the flow chart of operation is illustrated which determines and outputs it from T _o, continuously by performing this operation on the observed vibration waveform W (t), it is possible to know the time variation of T _o. For example, T
_{When o} is about 0.3 seconds, M is 40, K is 300, F
If the 100 Hz, _{T o} is output every 2 minutes,
1 day to obtain the variation waveform represented by T _o in data 720 points.

By computer simulation,
The result of estimating the natural period from the response vibration waveform of the system by applying an uncertain external force to the one-degree-of-freedom system (single resonance system) will be described below. FIG. 5 shows an acceleration vibration waveform (a) of an uncertain external force and its Fourier spectrum (b). The natural period of the system is 0.30 seconds, h (= 1 / 2Q) is 0.02, 0.0
4 and 0.08, the response waveform (observed vibration waveform) of the system when the uncertain external force shown in FIG. 5 was input was obtained, and the frequency distribution of the period was obtained from this response waveform.

FIG. 6 shows the Fourier spectrum (c) of the response waveform shown for comparison with the response waveform (a) and the frequency distribution of the period (b) when h is set to 0.02. FIG. 7 shows h
Is a response waveform (a), a frequency distribution of a cycle (b), and a Fourier spectrum (c) in the case where is set to 0.04. FIG. 8 shows the response waveform (a), the frequency distribution of the period (b), and the Fourier spectrum (c) when h is set to 0.08.

The above frequency distribution is expressed by (Equation 5) where n =
1, M = 40, F = 100, the cycle T was changed from 0.20 seconds to 0.40 seconds in increments of 0.01 second, and the number K of data of each frequency distribution was 300; (T = 0.20 and 0.40) are excluded. From these results, it can be seen that even in the case where it is difficult to detect with the Fourier spectrum (FIG. 7C and FIG. 8C), the natural period can be detected with high accuracy according to the frequency distribution of the period.

[0026]

According to the present invention, the natural frequency of a structure can be detected with high accuracy from the observed vibration waveform of the structure driven by wind, earthquake, microtremor or vehicle running vibration. It is suitable for capturing changes in natural frequency due to deterioration, and has the effect of enabling nondestructive inspection of strength deterioration by continuously measuring the natural frequency of a structure.

Further, the present invention has an effect that structural changes or abnormalities occurring in an operating power plant, an operating ship or an aircraft can be monitored by continuous measurement of a natural frequency.

[0028]

[Brief description of the drawings]

FIG. 1 is a diagram for explaining the operation of the present invention.

FIG. 2 is a diagram for explaining the operation of the present invention.

FIG. 3 is a diagram for explaining the operation of the present invention.

FIG. 4 is a diagram showing a flowchart of software for implementing the present invention on a computer.

FIG. 5 is a diagram showing an acceleration vibration waveform of an uncertain external force (a) and its Fourier spectrum (b).

FIG. 6 is a diagram showing a response waveform (a), a frequency distribution of periods (b), and a Fourier spectrum (c) when h = 0.02.

FIG. 7 is a diagram showing a response waveform (a), a frequency distribution of periods (b), and a Fourier spectrum (c) when h = 0.04.

FIG. 8 is a diagram showing a response waveform (a), a frequency distribution of periods (b), and a Fourier spectrum (c) when h = 0.08.

[Explanation of symbols]

 1 partial waveform reading block 2 waveform analysis block 3 frequency distribution creation block

Claims (1)

[Claims] 1. A means for dividing an observed vibration waveform of a structure vibrating due to an uncertain external force at short intervals to obtain a partial waveform; a means for approximating the partial waveform with a single periodic waveform; Using means for determining the frequency distribution of the cycle of the periodic waveform that approximates the series of partial waveforms obtained,
A structure using an uncertain external force, wherein a natural frequency of the structure is obtained from a cycle at which the frequency distribution is maximum, and strength deterioration of the structure is detected from irreversible fluctuation of the natural frequency. A method for detecting the strength deterioration of an object.